The plant circadian clock controls many aspects of growth and development, allowing an individual to adapt its physiology and metabolism in anticipation of diurnal and seasonal environmental changes. Circadian regulation of hormone levels and hormonal signalling modulates many features of development, including daily growth patterns and the breaking of seed dormancy. The clock also plays a role in seasonal day-length perception, allowing plants to optimally time key development transitions, such as reproduction. Moreover, the clock restricts (gates) the sensitivity of a plant's response to environmental cues, such as light and stress, to specific times of the day, ensuring that the plant can distinguish between normal fluctuations and longer-term changes. The central oscillator controls many of these output pathways via rhythmic gene expression, with several of the core clock components encoding transcription factors. Post-transcriptional processes are also likely to make an important contribution to the circadian regulation of output pathways. The plant circadian clock plays a role in regulating fitness, hybrid vigour and numerous stress responses. Thus elucidating the complexities of the circadian output mechanisms and their regulation may provide new avenues for crop enhancement.
Plants are sessile organisms, which as such are incapable of evading adverse environmental conditions. To compensate, plants exhibit a very high level of phenotypic plasticity, enabling their adaptation to variable growth conditions. In addition, they rely on their circadian clock in order to modify their physiology and metabolism in anticipation of predictable changes in environmental light and temperature conditions. Approx. 12% of Arabidopsis genes are controlled by the clock and exhibit circadian rhythmicity in constant light . The percentage of rhythmic genes increases to 89% under varied diurnal environmental cycles. The plant circadian clock plays an important role in photosynthesis by ensuring that expression of many of the genes involved in the light-harvesting reactions takes place at the optimal time of the day . Furthermore, it controls the degradation of starch during the night to ensure that reserves last until dawn . The clock also mediates the co-ordination of metabolic pathways for nitrogen assimilation and utilization .
Possessing a clock whose period matches the environmental cycle was shown to confer adaptive fitness in terms of growth rates and seedling survival . Under natural conditions, the clock is normally entrained to diurnal light–dark cycles; the timing of rhythms relative to the day–night cycle (commonly referred to as ‘phase') is determined in part by the free-running period in constant conditions . In hybrid and allopolyploid plants, increased chlorophyll, sugar and starch contents were found to correlate with altered epigenetic regulation of the core oscillator genes LHY, (LATE ELONGATED HYPOCOTYL) and CCA1 (CIRCADIAN CLOCK-ASSOCIATED 1), which resulted in altered amplitude of oscillations of their transcripts. This suggested that the increased growth vigour of these plants may be explained by modified circadian regulation of physiological and metabolic pathways . Thus developing a detailed understanding of circadian output pathway mechanisms may suggest new strategies for crop improvement by altering the timing of specific rhythms relative to cyclic environmental changes.
The oscillator mechanism
The mechanism of the central oscillator has been reviewed in detail elsewhere . Briefly, the plant clock is similar to that of animals, being composed of interlocked transcriptional feedback loops (Figure 1). However, the genetic components of these feedback loops are completely distinct. A core feedback loop comprises two MYB transcription factors, LHY and CCA1, with largely overlapping functions. These proteins are expressed in the morning and act to repress transcription of the TOC1 (TIMING OF CAB-1) gene, which encodes a pseudo-response regulator. LHY and CCA1 levels decline in the evening, allowing TOC1 expression to resume. This in turn results in activation of LHY and CCA1 transcription in the morning . A second feedback loop is mediated by PRR7 (PSEUDO-RESPONSE REGULATOR 7) and PRR9. Transcription of these genes is promoted by LHY and CCA1, and accumulation of their protein products subsequently results in down-regulation of LHY and CCA1 transcription . A mathematical model incorporating these two feedback loops, with a third mediated by TOC1 and an unknown component termed ‘Y', recapitulated most of the experimental data available . The nuclear protein GI (GIGANTEA) was suggested to function as a factor of ‘Y', as its temporal expression pattern matches that of the predicted component. The current model now requires revision, as the GI protein was recently shown to not act in a transcriptional feedback loop, but to modulate turnover of the TOC1 protein via its interaction with the F-box photoreceptor ZTL (ZEITLUPE). Binding of the rhythmically expressed GI protein to ZTL in the evening results in its stabilization. ZTL and GI then form a complex with TOC1 and target it for degradation by the proteasome (Figure 1). Although not essential for clock function, this process was shown to be important to maximize the amplitude of TOC1 protein oscillations . The pseudo-regulator protein PRR5 seems to carry out a dual function. It acts as part of feedback loop 1 to enhance TOC1 activity by promoting its phosphorylation, nuclear accumulation and recruitment to nuclear foci , and as part of loop 2 to mediate repression of LHY and CCA1 transcription in a manner similar to PRR7 and PRR9 .
The circadian oscillator is likely to comprise additional levels of regulation, as a number of further components have been identified whose function remains to be mapped to this network. These include CK2 (CASEIN KINASE 2) , the GARP transcription factor LUX [LUX ARRHYTHMO; also known as PCL1 (PHYTOCLOCK 1)] [16,17], the novel nuclear proteins ELF4 (EARLY FLOWERING 4), TIC (TIME FOR COFFEE), FIO1 (FIONA 1) and XCT (XAP5 CIRCADIAN TIME KEEPER) [18–20], the small GTPase LIP1 (LIGHT-INSENSITIVE PERIOD 1)  and two LWD (LIGHT-REGULATED WD) proteins, LWD1 and LWD2 . Mutations of ELF4 and LUX produce arrhythmic phenotypes, suggesting that these genes function close to or as part of the core mechanism of the oscillator. On the other hand, genes that only affect period length may be associated with processes that modulate clock function, but are not essential for circadian rhythmicity. For example, nitrogen metabolism can feed back on the function of the core oscillator . The phytohormone ABA (abscisic acid), whose expression and effect is modulated by the clock, can also affect its period. In constant light, application of ABA to plants causes a lengthening of the period of gene expression from the CCA1, AtGRP7 [Arabidopsis thaliana GLYCINE-RICH PROTEIN 7, also known as CCR2 (COLD, CIRCADIAN RHYTHM AND RNA-BINDING 2)] and CAB2/LHCB (light-harvesting chlorophyll a/b binding)1*1 promoters. Conversely, in the dark, the period of rhythmic expression from the AtGRP7 promoter is shortened in ABA mutants . Recent work showed that expression of the core clock gene TOC1 is acutely induced by ABA, providing a clue to the mode of action of this hormone. Furthermore, TOC1 binds the promoter of the ABAR (ABA-related) gene, and the resulting circadian expression of ABAR contributes to the gating of ABA effects on TOC1 . Thus ABAR and TOC1 function together in a regulatory feedback loop that is not essential for the function of the circadian oscillator, but may modulate it in response to developmental or environmental signals. Cytokinins have also been found to delay phase, whereas the addition of brassinosteroids can change circadian periodicity . Thus, multiple metabolic and hormone pathways can feed back into the circadian system to regulate the function of the clock. Indeed, a proportion of the mutations affecting circadian period may reflect alterations of these more peripheral processes.
Recent evidence indicates that the mechanism of the clock may vary between tissues and organs. The pseudo-regulator protein PRR3 was shown to function specifically in vascular tissue to regulate stability of the TOC1 protein . Furthermore, the circadian clock in Arabidopsis roots was shown to consist of a simplified version of the clock in shoots . In roots, two of the clock feedback loops are disengaged because the transcription factors CCA1 and LHY are unable to inhibit expression of TOC1 and GI. Under diurnal light–dark cycles, the shoot and root clock are synchronized through the action of an unknown metabolic signal.
Control of rhythmic gene expression
As transcriptional regulation forms the basis of the oscillator mechanism, the root of many circadian output pathways is expected to reside at the level of rhythmic gene expression. Several of the clock genes encode transcription factors and regulate the oscillatory expression of many downstream genes. For example, CCA1 and LHY are believed to repress expression of evening-specific genes by binding to the EE (Evening Element: AAATATCT; a sequence identified as being over-represented in sets of evening-specific promoters) . LHY and CCA1 may also promote transcription of morning- and midday-specific genes through interactions with the CBS (CCA1-binding site: AAAAATCT; originally identified in the promoter of midday-specific LHCB genes) [28,29]. Both CBS and EE elements are sufficient for cyclic transcription as synthetic promoters comprising multimerized copies of these sequences can confer rhythmic expression on luciferase reporter constructs [30,31]. Genome-wide ChIP-Seq (chromatin immunoprecipitation-sequencing) identification of in vivo binding sites for LHY indicated that more than 1000 promoters were recognized by this transcription factor, which represents approx. 3% of the genome (S. Adams, S. Veflingstad, S. Ott and I. Carré, unpublished work). Thus regulation by LHY alone may account for a significant fraction of the rhythmic gene expression observed in Arabidopsis. However, not all of the LHY target genes exhibited cyclic expression, and the phase of peak expression of rhythmic target genes ranged from late morning (ZT4) until late night (ZT20) (ZT is Zeitgeber time). Modulation of LHY activity by other transcription factors may account for this broad range of expression patterns. A number of other transcription factors and transcription-factor-binding sites have been shown to contribute to rhythmic gene expression (Table 1). In addition, changes in chromatin structure may act to modulate the timing and amplitude of circadian gene expression. For example, rhythmic transcription of the TOC1 locus is associated with rhythmic acetylation of histones at the promoter . Moreover, treatment of plants with Trichostatin A, an inhibitor of histone deacetylases, modified the phase and amplitude of TOC1 oscillations, resulting in an altered period length. The genome-wide contribution of epigenetic modifications to circadian-regulated gene expression remains to be determined.
Post-transcriptional processes also play a role in the regulation of gene expression. Expression of the glycine-rich RNA-binding protein AtGRP7 is regulated both by the clock and by various biotic stresses  and modulates tolerance to cold, drought and high salinity . AtGRP7 exhibits negative autoregulation by binding to its own pre-mRNA and promoting alternative splicing; the alternatively spliced transcript is then degraded. This negative-feedback loop acts downstream of the clock as part of a slave oscillator, signalling temporal information via the regulation of target transcripts . For example, during cold stress, AtGRP7 is involved in the transport of mRNA from the nucleus to the cytoplasm and may act as an RNA chaperone, modulating transcript folding to facilitate nuclear export .
Gating and anticipation of environmental responses
Plants are highly responsive to changes in environmental conditions. However, it is important to make the distinction between ‘normal' daily fluctuations and longer-term environmental changes. Thus one of the roles of the circadian clock is to minimize inappropriate responses by restricting sensitivity to specific intervals of the day (a phenomenon described as ‘gating'). For example, the clock modulates light responses to ensure that maximum sensitivity coincides with the middle of the day and that light signals perceived during the night have little or no effect. Expression of the LHCB genes was induced in response to light signals given during the subjective day, but minimal responses were observed during the subjective night . The circadian clock itself is modulated by light and this ensures its entrainment to light–dark cycles. This effect of light on the clock is also limited to specific times of the day. ELF3, a clock-associated protein, plays an important role in the gating of light responses, by attenuating light responses during the night . ELF3 modulates light input to the clock by regulating the proteasomal degradation of the GI protein . In the light, GI forms a complex with the F-box blue-light photoreceptor ZTL and promotes its accumulation [12,39]. ZTL in turn interacts with one of the core components of the oscillators, TOC1, and targets it for degradation . It is proposed that at night ELF3 acts as a substrate adaptor, allowing the ring-finger ubiquitin ligase COP1 (CONSTITUTIVE PHOTOMORPHOGENIC 1) to interact with GI and promote its degradation. The resulting destabilization of the ZTL protein would limit light input to the clock during the night .
The circadian clock also gates a plant's response to various stresses. In Arabidopsis, for example, cold treatments applied during the day (ZT4) induced cold responses more effectively than those during the night (ZT16) . Many clock-regulated genes (68%) were also identified as stress-responsive genes, suggesting that another important function of the clock is to enable anticipation of daily stress conditions, such as falling temperatures in the evening. In support of this hypothesis, disruption of circadian clock function in the arrhythmic prr9/prr7/prr5 mutant resulted in up-regulation of a significant number of cold-responsive genes. These plants also showed increased tolerance to various stresses, including cold, drought and high salinity . Furthermore, plants that overexpressed the TOC1 gene (TOC1-ox) or had reduced expression of the ABA receptor (ABAR-RNAi; RNAi is RNA interference) exhibited defective responses to drought, suggesting that the ABAR/TOC1 feedback loop described earlier is important to ensure survival in dry environments .
Rhythmic expression of the cold-inducible gene DREB1C (an early component of cold and dehydration signalling pathways, also known as CBF2) is mediated by the bHLH (basic helix–loop–helix) transcription factor PIF7 (phytochrome-interacting factor 7). PIF7 represses expression of DREB1C via binding to a G-box element within the promoter. Its effect is potentiated by physical interaction with the clock protein TOC1, resulting in rhythmic inactivation of DREB1C transcription . LHY and CCA1 may also contribute to cold responses, as bioinformatic studies showed that the EE sequence was enriched in the promoters of cold-induced genes . Mutational analysis of EEs and of shorter EEL (EE-like) motifs in the promoters of the circadian-regulated and cold-inducible genes COR27 and COL1 showed that these sequences were required for cold induction and that their effect was amplified by interactions with ABREL (ABA response element-like) sequences (ACGTG) . Statistical analyses indicated that combinations of EE, EE-like or CBS sequences with the ABREL were significantly enriched within sets of cold-regulated genes. Thus cold induction of gene expression may be mediated via the interaction of LHY and CCA1, or related MYB transcription factors, binding EE/EEL sequences, with bZIP (basic leucine zipper) transcription factors binding the ABREL. Combinations of EE, EEL or CBS sequences with a G-box motif (CACGTG) were also enriched within sets of cold-regulated genes, suggesting that LHY and CCA1 may also interact with bHLH transcription factors binding the sequence CANNTG to mediate cold responses .
These findings indicate intimate links between the clock-mechanism and stress-response pathways. Circadian regulation of stress-response genes may serve to minimize their deleterious effects on plant growth, and modulation of the timing of expression of these genes may offer more promising avenues for crop improvement than constitutive overexpression.
Hormonal control of plant growth and development
In addition to its role in tuning plant physiology to the environment, the circadian clock regulates growth and development through its effects on phytohormones. Under diurnal conditions, bioactive levels of auxin, gibberellins, brassinosteroids, ABA and ethylene accumulate specifically in the morning . The considerable overlap between sets of hormone- and clock-regulated genes suggests that oscillations of hormone levels may underlie rhythmic expression of many cyclic genes [46,47].
Many genes implicated in ABA synthesis are under clock control, with the majority showing peak expression during the morning . This is likely to mediate both the rhythmic production of ABA and, indirectly, the circadian regulation of ABA-responsive genes, since the expression of more than 40% of ABA-induced genes coincides with that of ABA-biosynthesis genes. Similarly, the production of ethylene is rhythmic, with levels peaking in the middle of the subjective day . The exact mechanism by which the clock controls cyclic ethylene production is unclear; however, several ethylene-synthesis genes exhibit circadian regulation and their peak of expression coincides with that of ethylene emission . Key components of the ethylene signalling pathway (EIN3 and EIL1) are also under clock control. As neither of these genes is regulated by ethylene, their circadian regulation must be independent of rhythmic ethylene emission. Thus the clock regulates hormone signalling at multiple levels. Genes involved in almost all aspects of auxin signalling display circadian regulation. Furthermore, transcriptional and growth responses to the application of exogenous auxin are gated by the clock, with maximum responsiveness in the subjective night . Temporal co-ordination of phytohormone transcript abundance may be mediated, at least in part, via the HUD (Hormone Up at Dawn) element (Table 1), a short DNA sequence motif (CACATG) that is over-represented in phytohormone gene promoters and, when multimerized, can confer time-of-day-specific expression on a reporter gene . In addition, the circadian-regulated transcription factor RVE1 (REVEILLE1) is essential for diurnal auxin rhythms and promotes the production of free auxin during the day through transcriptional activation of the auxin-biosynthetic gene YUCCA8 . Thus RVE1 provides a mechanistic link between the clock- and auxin-signalling pathways.
The circadian regulation of hormone-signalling components may allow the temporal integration of endogenous pathways, with external environmental cues to fine-tune development. For example, Arabidopsis seedlings exhibit rhythmic growth of their hypocotyls, with elongation peaking at subjective dusk in continuous light . However, under diurnal cycles, the peak of growth shifts towards dawn, due to interactions between circadian and light cues . Circadian and light-controlled patterns of phytohormone gene expression were shown to correlate well with these temporal patterns of hypocotyl growth. Peaks of phytohormone gene expression coincided with dawn under short-day conditions and with dusk in continuous light, resulting in the contrasting patterns of growth .
Seed dormancy allows plants to time their germination to correspond to suitable environmental conditions. The breaking of seed dormancy is dependent upon environmental factors, such as light and temperature. In addition, a process called dry after-ripening promotes the loss of dormancy and is thought to enable rapid germination following an extended drought period. Environmental cues have been shown to regulate seed dormancy and germination by modulating levels of two phytohormones with conflicting action. ABA acts to both establish and maintain seed dormancy and inhibit germination, whereas gibberellins promote the breaking of dormancy and subsequent germination . As several enzymes involved in gibberellin and ABA metabolism are controlled by the circadian clock, it was perhaps unsurprising to find that mutations that alter circadian rhythms also alter germination frequencies . A number of Arabidopsis mutants with abnormal circadian clocks (including lhy, cca1, gi, ztl and lux) showed environmental-sensing defects in seeds, resulting in altered levels of dormancy. A mechanistic link between the circadian oscillator and dormancy control was suggested by the observation that TOC1 can bind to a central regulator of dormancy, ABI3 (ABA-INSENSITIVE 3), in yeast two-hybrid assays . However, the process by which the clock affects seed dormancy is unclear. The clock appears to be arrested in an evening-like state in dry seeds, restarting when the seeds are imbibed. In response to imbibition, CCA1 expression was sharply up-regulated in non-dormant seeds, but not in dormant seeds. Distinct patterns of circadian-regulated gene expression followed .This suggests that signals modulating dormancy status act on an unknown factor that in turn regulates CCA1 expression at the time of imbibition to specify subsequent patterns of circadian oscillations and modulate hormone levels.
Responses to day length
The plant circadian clock also mediates perception of seasonal changes in day length in order to modulate flowering time and alter plant architecture.
The ability of plants to regulate the transition from vegetative to reproductive growth in response to day length allows them to initiate reproduction at the most favourable time of the year. The photoperiodic regulation of flowering has been covered in detail in recent reviews [55a,55b]. Thus, in brief, in Arabidopsis, the transcription factor CO (CONSTANS) plays a central role in the perception of day length and flowering. Expression of CO is finely tuned by the circadian clock, with RNA levels peaking 16 h after dawn . Under short-day conditions, expression of CO mRNA occurs after dark. The CO protein product is rapidly degraded in these conditions and flowering is inhibited (Figure 2A). The dark-dependent degradation of CO is regulated by COP1, which targets the protein to the proteasome . In contrast, under long days, peak expression occurs during daylight hours, when the activity of COP1 is inhibited by the action of the cryptochromes and by its exclusion from the nucleus (Figure 2B) . As a result, CO protein accumulates and promotes flowering via induction of the floral integrators FT (FLOWERING LOCUS T) and TSF (TWIN SISTER OF FT) [59,60]. Thus the photoperiodic regulation of flowering is mediated by the coincidence of an endogenous rhythm (CO) with an external signal (light). Another level of external coincidence is mediated by the blue-light photoreceptor FKF1 (FLAVIN-BINDING, KELCH REPEAT, F-BOX 1), which acts together with GI to promote the degradation of the transcriptional repressor CDF1 (CYCLIC DOF FACTOR 1) and enable CO transcription at the end of a long day . The CO–FT pathway is conserved in a wide variety of plant species and a functional CO homologue has been identified in Chlamydomonas, a unicellular photosynthetic green alga, suggesting that this pathway developed early in the chlorophyte lineage to regulate development in response to photoperiod .
Plant growth is also responsive to daylength. The bHLH transcription factors PIF4 and PIF5 allow adjustment of growth patterns to seasonal changes in photoperiod. Regulation takes place through an external coincidence mechanism similar to that described for the photoperiodic control of flowering. Transcription of the PIF4 and PIF5 genes is under circadian control, and stability of the protein products is modulated by light . Under short-day conditions, expression of PIF4 and PIF5 is restricted to the dark portion of the cycle. Both proteins accumulate to a high level, promoting hypocotyl elongation (Figure 2C). However, under long-day conditions, expression of the PIF4 and PIF5 genes coincides with light and their protein products fail to accumulate, resulting in a short-hypocotyl phenotype (Figure 2D) .
Conclusions and perspectives
Evidence so far suggests that the mechanism of circadian output pathways is largely transcriptional. Transcription-factor-binding sites that are associated with rhythmic gene expression are gradually being uncovered. A major challenge, however, will be to understand how such a great diversity of gene-expression patterns can be mediated by an oscillator composed of a small set of genes, and an even smaller set of transcription factors. Rhythmic hormone accumulation is likely to regulate an important fraction of the circadian transcriptome. Regulated protein turnover may represent a key regulatory mechanism, as this contributes both to the function of the oscillator and to the photoperiodic regulation of flowering. The F-box proteins ZTL and FKF1 are likely to target proteins other than TOC1 and CDF1 for ubiquitination and proteasomal degradation. Similarly, the ELF3 protein may direct proteins other than GI to the ubiquitin ligase COP1. Protein phoshorylation may also play a role. For example, the maize transcription factor Opaque 2 is expressed constitutively, but activated periodically through rhythmic phosphorylation . A small number of microRNAs display circadian regulation in Drosophila, suggesting a role in regulating transcript stability and translational efficiency . No such data are available for plants, although diurnal accumulation of four microRNAs has been reported in Arabidopsis . It also remains to be determined whether circadian oscillations of cytoplasmic calcium concentrations contribute to rhythmic changes in plant physiology.
The photoperiodic regulation of flowering time and that of hypocotyl elongation are mediated by similar mechanisms, involving rhythmic expression of one or more transcription factors and light-regulated protein turnover. This suggests that other photoperiod-sensitive aspects of development, such as germination, the formation of vegetative organs such as bulbs and tubers, and leaf abscission, may be regulated by similar mechanisms.
• The plant circadian clock is similar to animal clocks in that it is composed of interlocked transcriptional feedback loops. However, its genetic components are distinct.
• Immediate output pathways are mediated through transcriptional regulation by the master clock genes LHY/CCA1 and TOC1. Rhythmic protein degradation regulated by GI, ELF3 and the blue-light photoreceptor, F-box protein, ZTL may also mediate circadian output pathways.
• Downstream events include the rhythmic regulation of phytohormone synthesis and signalling.
• Circadian gating of light responses is mediated, at least in part, by the ELF3-mediated targeting of the GI protein for ubiquitination and proteasomal degradation.
• Day-length regulation of growth and flowering are mediated by similar external coincidence mechanisms involving the rhythmic accumulation of a light-labile protein that only accumulates to active levels under short-day conditions, or of a dark-labile protein that only accumulates to active levels under long-day conditions.
S.A. is supported by BBSRC grant BB/F022832/1 (to I.A.C.).
- © The Authors Journal compilation © 2011 Biochemical Society